802.11ay Wi-Fi: “It’s going to be a very scalable spec”
- 28 March, 2017 21:00
Peraso Technologies, one of the early players in 802.11ad (WiGig) gigabit wireless networking, is also readying for enhanced Wi-Fi technology called 802.11ay that promises higher speeds and greater reach.
The Toronto-based company, whose semiconductors enable smartphones, tablets and other devices to exploit unlicensed 60 GHz wireless technology, is bullish on 11ay to support an even wider range of products than 11ad.
(See Also: Our FAQ on 802.11ay)
Peraso Co-Founder and Senior VP of Product Development Brad Lynch spoke with Network World recently about the promise of 11ay.
How does 11ay compare to 11ad?
It’s almost like if you could compare 11ac to 11g — it’s almost that big of leap in terms of performance and specs between the two. 11ay as the specs are being developed is really allowing for a wider range of products than you’d get with ad, which has one set of data rates that everyone supports [most 11ad products support a 4.6Gbps mode]… ay has a lot more parameters to play with in channel bonding, MIMO (multiple input, multiple output) and features at the MAC level to allow a far greater range of performance and products… It’s going to be a very scalable spec, much more so than 11ad.
What speed and distance parameters are we talking about with ay
A 20 to 30Gbps speed — the basic PHY rate if you will is probably going to be around 30Gbps — and a range of say 10 meters between any two ay devices, maybe not at the full 30Gbps but still 10Gbps-plus. What you’ll start to see with ay is the option of instead of just having two basic devices maybe you’ll have a high performance access point for an office, so now instead of 10 meter range you’re going to get 30 meters of range and cover a fairly large cubicle area. Instead of just being single-user MIMO capable, which is what you get with a 30Gbps link, you could go with multi-user MIMO so you could have 2 or 4 devices simultaneously doing that 30Gbps. In 11ay, because they’re putting in MIMO from the ground up, once you turn it on all devices will be able to do that. So even if you have a single-user MIMO device it will still support a multi-user MIMO peer, which is something you don’t always see with 11ac today. There’s tremendous opportunity for performance in ac but how can you have one slow device that bogs everything down? I think you’ll see a lot less of that with 11ay.
Is it appropriate to think of 11ay as an extension of 11ad?
Absolutely, in the same way 11ac was an extension of 11n or 11a and g. 11ad will be a subset of ay. All the basic principles of how the networks form is the same as with ad. What’s being done is adding these much greater PHY rates on top of it.
Speaking of 11ad, how is that market going?
Definitely this year feels like the year that it’s going to ramp up. At the big shows for us like CES and Mobile World Congress there’s been a big change from people saying "That technology’s interesting, come back to us in 6 months when you’ve gotten here or here," to more of them saying, “OK, we’ve got to get started with design because we feel this is going to happen in the first or second half of the year.” I think 2017 will really be a year when people start to ship a lot more 11ad products [Editor’s note: Routers from companies like Netgear have hit the market, but there aren’t a lot of devices to connect to them, Lynch says]. One of the factors is that 11ac has kind of reached the limits of its performance, especially on the consumer side. If you need true gigabit wireless networking for your application, ad is the only way you’re going to get there… Late last year the Wi-Fi Alliance completed its first certification program for ad, so that’s obviously a big step that I know I can go out and buy a router or laptop and they’ll actually work together.
With 11ad still in a formative period, why should enterprises be thinking of ay?
There will be awareness that 11ad is just the first gear of what you’re going to get with millimeter wave products. If you buy an ad product today it’s going to work with an ay product that comes out two years from now. It makes sense to get comfortable with the realities of deploying it, particularly in the enterprise. One of the things that ad does, and to some extent it’s a benefit, is that it doesn’t go through walls very well if at all. So that means your deployment model has to change a little bit. If you have offices and put an access point in the ceiling that’s great but it’s not going into someone’s office at that point so you’re either going to need to have a small access point in there or cover that office in some other way. So you’re going to deploy access points in a denser formation. As ay rolls out, what it will bring in an enterprise scenario if ad doesn’t quite do it for you, is it will finally be the technology that lets you snip that Ethernet cord. You no longer have to run Ethernet cables to everyone’s desk… there’s enough wireless bandwidth in ay.
I’ve heard 11ay described as a good backhaul option as well…
That’s right. Outdoor access is the other area we’re seeing. Google made a lot of waves a few years ago with its fiber initiative and they were just going to run fiber everywhere, but they ran into the realities of economics…it’s just too costly to dig trenches and run fiber up poles. One of our customers uses 11ad to let people deploy low cost gigabit wireless links. 11ay is really going to up that rate further to in excess of a gigabit per second and bring that cost down even more because I have 30Gbps at a node from which I can serve up a gigabit to 30 people.
What’s the 11ay timeline look like?
The [IEEE Task Group for ay’s] first draft is targeted for later this year. My sense is the progress being made is very good… because there are companies shipping ad products there won’t be nearly the long lag between the ay standard being done and the products rolling out that you saw with ad. If the first draft does get completed later this year, I think that gives an opportunity for products to roll out within 12 months of that in a pre-ay form, much like you saw with ac and 11n. Simply because there are so many applications that can make use of the technology, including in the VR space, where you’ll be able to cut that cord that you see with an HTV Vive or Oculus Rift or PlayStation VR today, or if you want to do anything with AR and VR in the enterprise.
How else might 11ay apply to the enterprise?
Cloud applications inherently demand a lot of bandwidth and to get a good experience you need low latency and high performance. A lot of the easy applications have been moved to the cloud; the next [batch] is demanding a lot more bandwidth and lower latency, and ay will allow people to do more of that.